Page 119 -
P. 119

96     4. Families of Elliptic Curves and Geometric Properties of Torsion Points

        (5.1) Invariants of the Curve E[a, b]. For the curve E[a, b]definedby

                                       3
                                  2
                                             2
                                 y = x + ax + bx
        the following hold:

                             2                 5         3
                    c 4 = 16 a − 3b ,    c 6 = 2  9ab − 2a  ,
                                                       2
                                              c 3      a − 3b   3
                         4 2  2                4    8
                      = 2 b  a − 4b ,     j =    = 2           .
                                                         2
                                                     b 2  a − 4b
        The two basic special cases are:
                 3
         (1) j = 12 if and only if a = 0 where we abbreviate the notation E[0, b] = E[b]:
             2
                 3
            y = x + bx.
                                           2
                                   2
         (2) j = 0ifand onlyif3b = a = (3c) for characteristic unequal to 3. This is
                           2        2    3      2     2            2    2
            the curve E[3c, 3c ] where y = x + 3cx + 3cx x,or E[3c, 3c ]: y =
                      3
                                     2
                                          3
                  3
                                              3
            (x + c) − c . It has the form y = x − c after translation of x by c.
           Before giving the formulas for the isogeny, we observe that the function h(a, b) =
               2                        2                         2     2
        (−2a, a − 4b) when iterated twice is h (a, b) = (4a, 16b) since (−2a) − 4(a −
                               2
                                     2
        4b) = 16b. This means h : k → k is a bijection for char(k)  = 2, and it is its own
        inverse up to the division of the coordinates by a power of 2, namely
                                              2
                           h −1 (a, b) = (−a/2,(a + 4b)/16).
                       2
                                                          2
        Finally, note that a − 4b is the discriminant of the quadratic x + ax + b.
        (5.2) Formulas for the 2-Isogeny. The 2-isogeny with kernel {0,(0, 0)} is given by
                            2
        ϕ : E[a, b] → E[−2a, a − 4b] where
                            2      2
                           y   y x − b              b        b
                 ϕ(x, y) =   ,           = x + a +   , y 1 −      .
                           x 2    x 2               x        x 2
                                              2
        The formula for the dual isogeny ˆϕ : E[−2a, a − 4b] → E[a, b]isgiven by
                             2
                            y    y     2     2
                  ˆ ϕ(x, y) =  ,     x − a − 4b
                            4x 2  8x 2
                                        2                2
                            1          a − 4b    y     a − 4b
                        =      x − 2a +        ,    1 −          .
                            4             x      8        x 2
           Note that ˆϕ is given by the same formulas as ϕ up to powers of 2 in the coordi-
                                                        2
        nates which reflects the fact that the dual ˆϕ maps E[−2a, a − 4b]to E[a, b] while
                                                                          6
                       2
        ϕ maps E[−2a, a − 4b]to E[4a, 16b]. Multiplying the equation for E[a, b]by2 ,
        we see that (x, y) → (4x, 8y) is an isomorphism of E[a, b] → E[4a, 16b]. Hence
        any property that holds for ϕ will also hold for the dual example an easy calculation


        shows that ϕ(x, y) = (x , y ) satisfies
   114   115   116   117   118   119   120   121   122   123   124